Agreed. One of the problems with meta analyses, wh...

2011-12-16T21:07:34.795+10:00

Agreed. One of the problems with meta analyses, which I'm guilty of, is an uncritical over-reliance on them. That is, we assume the meta analysis is of a high, unbiased, quality. Ironically, one thing I've noticed about meta analyses in general is that many of the studies cited are criticised for being of poor quality - but how often do we independently query the quality of the meta analysis?

All that being said, I nevertheless agree with Pasher et al's, conclusion, that there is overall no confirming evidence of the efficacy of learning styles. And in order to advocate learning styles as an educational intervention we would need conclusive evidence of its efficacy.

I'm no educational theorist but I do know a th...

2011-12-16T16:01:40.352+10:00

I'm no educational theorist but I do know a thing or two about statistics and a thing or two about meta studies and reviews.

So my bullshit detector sounded alarms when I noticed that the sample size in the positive reporting paper by Sternberg et al. was highlighted in order to attack methodology, but was entirely absent in the description of negative reporting papers by Massa and Mayer and Cook et al.

And a quick glance at the originals showed why. The sample size was small. Small enough to have a standard deviation that could mask any real results. No matter how good the method and how good the testing a sample size of 60 is unlikely to be sufficient.

Which isn't to say anything conclusive has been show one way or another. But it does give an idea of what Pashler et al. were setting out to show and makes me quite suspicious of their paper.

Comments on The Skeptic's Field Guide: Three ways of examining brain claims Part Three - ...

Agreed. One of the problems with meta analyses, wh...

I'm no educational theorist but I do know a th...